Monday, January 26, 2015

Econometrics and Kung Fu

What do econometrics and Kung Fu have in common?

(This post has been updated to include links to teaching resources at the end.)

Tomorrow I'm giving the first lecture in my undergraduate econometrics course.  I'm excited to be teaching this important class for the first time.  I'm also excited to have the chance to use, for the first part of the class, an excellent new book that I think will motivate students as well as serve as a valuable resource.

Last month, Joshua Angrist and Jรถrn-Steffen Pischke published their latest econometrics book, Mastering 'Metrics.  From what I've read so far, this book will be more accessible to the beginning student than their earlier book, Mostly Harmless Econometrics.  They begin their most recent book with a passage from a famous scene in the 1972 pilot of the TV series Kung Fu:
Master Po: Close your eyes. What do you hear?
Young Caine: I hear the water, I hear the birds.
Po: Do you hear your own heartbeat?
Caine: No.
Po: Do you hear the grasshopper which is at your feet?
Caine: Old man, how is it that you hear these things?
Po: Young man, how is it that you do not?
What's the connection between their book and Kung Fu?   They describe this on pages xi-xii of the Intro: 
 There is a mystical aspect to our work as well: we're after truth, but truth is not revealed in full, and the messages the data transmit require interpretation.  In this spirit, we draw inspiration from the journey of Kwai Chang Caine, hero of the classic Kung Fu TV series.  Caine, a mixed-race Shaolin monk, wanders in search of his U.S.-born half-brother in the nineteenth century American West.  As he searches, Caine questions all he sees in human affairs, uncovering hidden relationships and deeper meanings.  Like Caine's journey, the Way of 'Metrics is illuminated by questions.
I'm guessing most students have not seen this series.  Therefore, tomorrow I plan to show the scene from which the passage above was taken, and also a bit before and after to provide context (fellow instructors may wish to know I plan on starting at 17:20 and going for about 6 and a half minutes.)  I think this will be class time well spent as it will not merely allow me to entertain the students, but will also allow me to capture their interest and teach them valuable lessons.

This scenes should be of interest for several reasons.  First, we see Caine on his first day on the job in the U.S., working with Chinese workers on the construction of a rail road.  Some of these scenes were filmed at Vasquez Rocks in the Sierra Pelona Mountains in northern Los Angeles County.  Incidentally, this site is also close to potential routes of California's High Speed Rail project.

On a more pedagogical level, the segment ends with young Caine's recognition that learning a difficult subject takes time.  I think many students become frustrated with econometrics because of this.  It should be reassuring to know that learning it is difficult for everyone, even those who go on to become "masters". 

Both the Introduction and Chapter 1 of Angrist and Pischke's book are available for free download from the publisher's website.  These are valuable, free resources that I will use in my course.  Perhaps most useful for us will be the Appendix to Chapter 1 (pages 34-46 here) which provides a more readable, focused and concise review of statistics than can be achieved by assigning the lengthy (and somewhat encyclopedic) chapters from their main textbook.  Students should be happy that I am able to replace about 100 pages of traditional textbook reading with these enjoyable and effective resources!

The first Chapter 1 covers Randomized Trials.  Randomized Trials are one of Angrist and Pischke's "...Furious Five of econometric research..." which are "...random assignment, regression, instrumental variables, regression discontinuity designs, and differences-in-differences..." (p. xiv).

Although I had always planned to discuss "Idealized Experiments" on the first day (a topic also covered in Chapter 1 of their main text) I'm covering them in much more detail at the start of the semester than I otherwise would have, thanks to the free resources offered by Angrist and Pischke, and Princeton University Press.

But beyond the zero cost, covering experiments first has an added pedagogical virtue in that it can be combined with an exercise that requires the students to conduct hypothesis tests, using data from an experiment that they generated through a quiz they took during the introductory meeting.  Not only should this help them understand the mechanics of hypothesis testing, but by having them create data from answers to quiz questions, they will see how survey design, data entry and formatting (or "cleaning") can be messy processes.

In the spirit of contributing some educational public goods, I plan to post some of the assignments I create for this course here and potentially to other places, so check back later, email me or comment here until I can add these links.

I have decided not to require their Chapter 2 on regression, because at only 50 pages I think the pace will be too quick for most students in an introductory course, where I plan to cover regression basics in detail.  Requiring it or other chapters on top of already extensive readings from the main textbook might be too much of a burden for many students.  However this doesn't mean I wouldn't recommend that a student who can handle the work load purchase and read this book.  Rather, I think Mastering 'Metrics would serve as a valuable complement to our main textbook.  Indeed, I'll be reading this book carefully myself this semester, and looking for ways to further incorporate it into our curriculum.

Update (2/3/2015):

As promised, I'm posting the teaching resources I've developed for this course so far.  Related to my post, I mentioned one of my assignments requires the students to conduct hypothesis tests, using data from an experiment that they generated through a quiz they took during the introductory meeting.  This is the Lab Assignment for Week 2, and it uses the Data in the worksheet "syllabus_quiz".  The students took the syllabus quiz, which is available under Quizzes, on the first day of class, after spending 10 minutes reading the Syllabus.

This syllabus quiz contained an experiment in priming, where half of the students had a quiz where the last two questions were: (9.) Is the population of Istanbul greater or less than 1.4 million? ______  and (10.) What do you think is the population of Istanbul?  Meanwhile, the other half of the students had a question where 1.4 million was replaced by 14 million in question 9, but was otherwise identical in all respects.  Given the best estimate of the population of Istanbul is much closer to 14 million, we could describe students in the former group as being in the “treatment” group (because we tried to manipulate this group), and the students in the latter group as being in the “control” group (because we told them the truth.)

Lab Assignments

I am updating some of these files periodically.  Comments are much appreciated!


  1. Awesome, thanks Matt! We'll add your resources to the web page - Master Joshway

  2. This comment has been removed by a blog administrator.

  3. Nice Blog! Experience the Best Assignment Writing Services at Assignment Help Sydney Australia

  4. Get professional assignment help services from Student Assignment Help. We have several qualified experts and researchers who holds expertise in different subjects who are always ready to work on the student request who asking Write my assignment for me.

  5. ABAssignmentHelp has been the best name in the UK in recent years. We generally work hard to give assignment help services according to the prescribed measures and principles that can help students achieve the best evaluation. We are providing the best writing services at an affordable price.


  6. A high-level post with a piece of knowledgeable information.Thank you for sharing such information.
    if you need any academic level Assignment Help at reliable quality with better work.
    for more: WhatsApp or call +61 2 80113341